The most immediate way would be to simply create some custom functions
based on the library ones. With big datasets though, you would
probably get better performance by using the C++ calculator available
at:
https://github.com/jeetsukumaran/possum
Either way, if you submit a feature request on GitHub, I'll see about
working on this on the next release.